In EM.group, there is EMhistory matrix that I think is very useful and could be part of a fitted object produced by mirt. For some SE.types, it is already returned by EM.group, but then it is sadly not carried any further (from what I see). It would be nice to have it for all EM fits somewhere in fit@Internals.
Seems reasonable to me. You can now add technical = list(storeEMhistory=TRUE) during estimation, and extract the history via extract.mirt(mod, 'EMhistory')
In
EM.group
, there isEMhistory
matrix that I think is very useful and could be part of a fitted object produced bymirt
. For someSE.type
s, it is already returned byEM.group
, but then it is sadly not carried any further (from what I see). It would be nice to have it for all EM fits somewhere infit@Internals
.Many thanks for considering the feature!